Our objective is to determine the major factors that influence the course of drug use disorders in non- clinical populations. The majority of people who have drug disorders do not enter the formal treatment system;nonetheless, a number achieve "natural recoveries." Mutual-help groups such as NA/CA may be important, but social support, social resources, comorbid psychopathology, and stressful events may also be influential. While there is an increasing body of research on natural recovery, most studies have been cross- sectional or longitudinal with long intervals between assessments. Such designs tell us little about how dynamic factors influence the course of drug use disorders. For other psychiatric disorders, naturalistic longitudinal studies in which subjects are followed intensively over time have provided insights into predictors of remission or exacerbation of these disorders, as well as a wealth of data on how people interact (or, often, fail to interact) with health care providers. The time is ripe for drug use disorders to be studied in a similar manner. The purpose of this pilot project is to provide key data necessary for the design of a larger and longer-term naturalistic study. We will recruit 60 clients from the community rather than from clinics, and will follow these cases intensively for one year. We will use instruments adapted from successful longitudinal studies of Axis I disorders to track psychiatric status, life events, and treatment dynamically. We will use standard timeline measures to record drug outcome dynamically;other potential predictors/mediators to be measured are behaviors and attitudes related to mutual-help attendance and affiliation, social support and resources, self-efficacy, and motivation. Successful design of a high-intensity longitudinal study, however, depends on factors such as frequency of comorbid disorders, rates of remission, relapse, and stressful events, and rates of treatment. The specific aims of this pilot study are to test recruitment and assessment procedures, and to estimate rates of key events such as return to use of cocaine to support power estimates for the study to come. Because longitudinal data give us an opportunity to study sequences of events over time, they provide opportunities to test theoretical predictions about the mechanisms of action of variables such as mutual help and social support, information that could help us design better treatments. Furthermore, a better understanding of the factors that lead people with drug problems in the community to choose, or not choose, formal treatment support and/or informal support systems, may lead us toward better systems for identifying and aiding persons at high risk. Thus, an improved understanding of the biological, psychological, and social forces operating in the community will help us design future intervention systems that can better target persons at high risk for preventive interventions, and be better tailored to take advantage of positive resources available to individual patients. PUBLIC HEALTH RELEVANCE: Most people who qualify for a drug abuse or dependence diagnosis do not use professional treatment for this condition;nonetheless, many are able to overcome the disorder and lead normal lives. We know very little about how people achieve remission without resorting to professional treatment;such information will be valuable in devising interventions for persons with drug use disorders that can take advantage of the strengths that people bring with them, and in coping with the risk factors that can lead to deterioration. This study represents the beginning of a research program on recovery and relapse in drug use disorders in the community.